Simulation - The Challenge for Data Science

Explore the challenges and opportunities in data science simulation models through this insightful lecture from the Alan Turing Institute. Delve into the limitations of machine learning and the potential of simulation models for addressing complex problems like policy analysis. Learn about the difficulties in parameter estimation and initialization for accurate simulations, and discover a new method inspired by meteorology for solving these issues. Gain valuable insights into the fusion of simulation science and machine learning, and understand why these problems are at the forefront of data science. Examine the application of agent-based modeling in economics and finance, and grasp the importance of simulation science in the future of data analysis. Benefit from the expertise of Doyne Farmer, a renowned researcher in complexity economics and financial instability, as he shares his perspectives on advancing simulation science to match the utility of machine learning.

Introduction
Trading in Markets
Background Comment
Why Simulation
Machine Learning
AgentBased Modeling
Traditional Economic Models
Closed Form Solutions
AgentBased Models
Advantages of AgentBased Models
Challenges of AgentBased Models
Design Philosophy
Housing Markets
Challenges
Parameter estimation
Timeseries forecasting
Snapshot
Weather Prediction
Conclusion